A Whole-local Damage Identification Method Of Cable-stayed Bridge Based On Wavelet Transformation

Under the influence of design,construction and external environment of normal operation,the stay-cables which are the main load-bearing component of the cable-stayed bridge will suffer stress relaxation,corrosion,partial broken wire and other damages,and the whole cable-stayed bridge will also suffer damage.These damages pose potential threat to the safety of cable-stayed bridge structures.Therefore,it is of great significance to identify the damage of the whole structure and the key parts of the cable-stayed bridge.In this paper,the wavelet transformation method was used to identify the damage status of the whole cable-stayed bridge and detect the number of broken wires in the local stay-cables.The research is divided into three parts which are theory,numerical simulation and experiment.First,by wavelet transformation of the response signal,the time-varying frequency of the structure was identified by the maximum value of the module of the wavelet coefficient.The time-varying mode shape of the structure was obtained by normalizing the ratio of the module of the wavelet coefficients.The damage of the structure was identified by the wavelet transformation of the variance of mode shape.This method was applied to identidy the damage of the cable-stayed bridge after collision.The results showed that the proposed method could not only accurately identify the time of collision,but also accurately identify the damage location and damage extent of the structure.Second,to identidy the damage of the local stay-cables which are the main loadbearing component of the cable-stayed bridge,the finite difference model of the stay-cables was established.A method to identify the damping ratio of the stay-cables based on wavelet transformation was proposed.The damping ratio of the stay-cables was identified by the differential of the logarithm of the modulus of the wavelet coefficients and the equations of the cable frequency.The proposed method was applied to a numerical example of Ganjiang Bridge and a full-scale experiment of stay-cables,the results showed that the proposed method can accurately identify the damping ratio of stay-cables and the installation position of the electromagnetic inertial damper of the stay-cables.Third,aiming at the problem that the inner steel wire rope of the stay-cables is easy to rust and the excessive noise of the MFL signal affects the effective information of the signal itself,a surface defect parameter identification method of the stay-cables based on magnetic field leakage wavelet transformation was proposed.An open loop permanent magnet magnetization model of the defective cable based on ANSYS was established.By setting different defect parameters,the defect length was identified by peak peak distance of the radial component of the denoising MFL signal in circumferential direction,the defect width was identified by peak peak distance of the radial component of the denoising MFL signal in axial direction,the defect depth was identified by peak peak value of the radial component of the denoising MFL signal in axial direction.Final,aiming at the problem that the internal steel wire rope of the stay-cables gradually expands and forms the broken wire,a method based on magnetic flux leakage wavelet neural network was proposed to identify the number of broken wires of steel wire rope in stay-cables.This method was applied to an experiment which was designed to detect the magnetic flux leakage of cable.The influences to the peak peak value of the MFL signal by the number of broken wires,the diameter of wire rope,the detection speed and the lifting value were investigated.The number of broken wires of steel wire rope in stay-cables were accurately identified by using this method.